Method, device and system for constructing user voiceprint model

The invention discloses a method, a device and a system for constructing a user voiceprint model. The method comprises steps: voice information inputted by the user is received; whether preset key words meeting modeling requirements are carried in the voice information is judged; if yes, a voice segment corresponding to the preset key words meeting the modeling requirements is captured from the voice information; and by using the voice segment, a voiceprint model is constructed for the user. When the technical scheme provided by the invention is applied, in an application scene of randomly receiving the voice of the user, construction of the user voiceprint model can be completed, and use by the user is facilitated.